Abstract
Parsimony haplotyping is the problem of finding a set of haplotypes of minimum cardinality that explains a given set of genotypes, where a genotype is explained by two haplotypes if it can be obtained as a combination of the two. This problem is NP-complete in the general case, but polynomially solvable for ðk; lÞ-bounded instances for certain k and l. Here, k denotes the
maximum number of ambiguous sites in any genotype, and l is the maximum number of genotypes that are ambiguous at the same site. Only the complexity of the ð; 2Þ-bounded problem is still unknown, where denotes no restriction. It has been proved that ð; 2Þ-bounded instances have compatibility graphs that can be constructed from cliques and circuits by pasting along an edge. In this paper, we give a constructive proof of the fact that ð; 2Þ-bounded instances are polynomially solvable if the compatibility graph is constructed by pasting cliques, trees and circuits along a bounded number of edges. We obtain this proof by solving a slightly generalized problem on circuits, trees and cliques respectively, and arguing that all possible combinations of optimal solutions for these graphs that are pasted along a bounded number of edges can be enumerated efficiently.
maximum number of ambiguous sites in any genotype, and l is the maximum number of genotypes that are ambiguous at the same site. Only the complexity of the ð; 2Þ-bounded problem is still unknown, where denotes no restriction. It has been proved that ð; 2Þ-bounded instances have compatibility graphs that can be constructed from cliques and circuits by pasting along an edge. In this paper, we give a constructive proof of the fact that ð; 2Þ-bounded instances are polynomially solvable if the compatibility graph is constructed by pasting cliques, trees and circuits along a bounded number of edges. We obtain this proof by solving a slightly generalized problem on circuits, trees and cliques respectively, and arguing that all possible combinations of optimal solutions for these graphs that are pasted along a bounded number of edges can be enumerated efficiently.
Original language | English |
---|---|
Pages (from-to) | 234-247 |
Journal | Ieee-Acm Transactions on Computational Biology and Bioinformatics |
Volume | 12 |
Issue number | 1 |
DOIs | |
Publication status | Published - 1 Jan 2015 |